Malvern Instruments Limited

Laser diffraction gives the end-user would be to measure in liquid

BASIC PRINCIPLES OF PARTICLE Equivalent spherical diameter of

TECHNICAL PAPER
the following advantages: suspension (aqueous or organics) Enigma Business Park cylinder 100 x 20µm
● The method is an absolute one
set in fundamental scientific ●
for the reasons discussed above.
The entire sample is measured.
Grovewood Road
Malvern SIZE ANALYSIS Imagine a cylinder of diameter
D1 = 20µm (i.e. r=10µm) and
principles. Hence there is no Although samples are small Worcs Written by Dr. Alan Rawle,Malvern Instruments Limited, Enigma Business Park, Grovewood Road, height 100µm.
WR14 1XZ Malvern, Worcestershire, WR14 1XZ, UK. Tel: +44 (0)1684 892456 Fax: +44 (0)1684 892789
need to calibrate an instrument (4-10g for dry powders, 1-2g for There is a sphere of diameter, D2
U.K. What is a Particle? which has an equivalent volume to
against a standard – in fact there is suspensions typically) and a
Tel: +44 (0)1684 892456 the cylinder. We can calculate this
no real way to calibrate a laser representative sample must be This may seem a fairly stupid question to ask! However, it is
Fax: +44 (0)1684 892789 diameter as follows:
diffraction instrument. obtained, all the sample passes fundamental in order to understand the results which come from
through the laser beam and various particle size analysis techniques. Dispersion processes and Volume of cylinder =
Equipment can be validated, to
Malvern Instruments S.A.
confirm that it is performing to diffraction is obtained from all the shape of materials makes particle size analysis a more complex 2 3
Parc Club de L’Université
matter than it first appears. πr h = 10000 π (µm )
certain traceable standards. the particles.
30, Rue Jean Rostand
● A wide dynamic range. The best ● The method is non-destructive and 91893 Orsay Cedex edge or to a diagonal. With our
laser diffraction equipment allows non-intrusive. Hence samples can France
High
Sphericity matchbox there are a number of Volume of sphere =
the user to measure in the range be recovered if they are valuable. Tél: +33 (1) 69 35 18 00 Medium properties of it that can be described
Sphericity 4 πX 3
from say 0.1 to 2000 microns. ● A volume distribution is generated Fax: +33 (1) 60 19 13 26 by one number. For example the
3
Smaller samples (1nm – 1µm) can directly which is equal to the
Low
Sphericity
weight is a single unique number as
be measured with the photon Malvern Instruments Inc. is the volume and surface area.
weight distribution if the density is Very Angular Sub- Sub Rounded Well Where X is equivalent volume radius.
correlation spectroscopy technique 10 Southville Road Angular Angular Rounded Rounded So if we have a technique that
constant. This is the preferred dis- measures the weight of the
as long as the material is in Southborough Properties:
tribution for chemical engineers. V = Volume matchbox, we can then convert this X=
3
3V = 0.62 3 V
suspension and does not sediment. MA 01772 W = Weight
weight into the weight of a sphere, 4π
● The method is rapid producing an d min.
S = Surface Area
● Flexibility. For example it is U.S.A. A = Projected Area remembering that…
answer in less than one minute. d max. R = Sedmentation Rate
possible to measure the output Tel: +1 (508) 480-0200
This means rapid feedback to 3
Fax: +1 (508) 460-9692 3OOOOπ = 3 7500 = 19.5µm
weight = 4 πr .p
Figure1 3
from a spray nozzle in a paint operating plants and repeat X=
3 4π
booth. This has been used by analyses are made very easily. Malvern Instruments GmbH The Particle size conundrum
nozzle designers, to optimise the
● Highly repeatable technique. Rigipsstraße 19 Imagine that I give you a matchbox and calculate one unique number
viscosity, ∆P and hole size and (2r) for the diameter of the sphere of D2 = 39.1µm
This means that the results can be 71083 Herrenberg and a ruler and ask you to tell me the
layout, in order to get correct size of it. You may reply saying that the same weight as our matchbox.
relied on and the plant manager Germany
droplet size. This has found the matchbox is 20 x 10 x 5mm. You This is the equivalent sphere theory. The volume equivalent spherical
knows that his product has Tel: +49 (0) 7032 97770
extensive application in the cannot correctly say "the matchbox is We measure some property of our diameter for a cylinder of 100µm
genuinely changed and that the Fax: +49 (0) 7032 77854
agricultural and pharmaceutical 20mm" as this is only one aspect of particle and assume that this refers to height and 20µm in diameter is
instrument is not "drifting". a sphere, hence deriving our one
industries. For further informa- its size. So it is not possible for you around 40µm. The table below indi-
● High resolution. Up to 100 size Malvern Instruments (Asia Pacific) unique number (the diameter of this
tion the reader is referred to to describe the 3-dimensional cates equivalent spherical diameters
classes within the range of the Penthouse 21-11 matchbox with one unique number. sphere) to describe our particle.
References 4,5 and 6. There is of cylinders of various ratios. The last
system can be calculated on the The Boulevard Obviously the situation is more This ensures that we do not have to line may be typical of a large clay
now an ASTM standard for sprays Mid Valley City
Malvern Mastersizer. difficult for a complex shape like a describe our 3-D particles with three particle which is discshaped. It
using laser diffraction. Lingkaren Syed Putra or more numbers which although
grain of sand or a pigment particle in would appear to be 20µm in diame-
● Dry powders can be measured References 59200 Kuala Lumpur a can of paint. If I am a Q.A. more accurate is inconvenient for ter, but as it is only 0.2µm in thick-
directly, although this may result 1. Paint and Surface coatings – Malaysia Manager, I will want one number management purposes. ness, normally we would not consider
in poorer dispersion than using a theory and practice; Ed. R. Lambourne Tel/Fax: 603 2283 6139 only to describe my particles – I will We can see that this can produce this dimension. On an instrument
liquid dispersing medium. Ellis Horwood Ltd. 1993. need to know if the average size has some interesting effects depending which measures the volume of the
ISBN 0-13-030974-5PGk increased or decreased since the last on the shape of the object and this is particle we would get an answer
However, in conjunction with a Malvern Instruments Nordic AB
2. Particle Size Measurement;T. Allen, Box 15045, Vallongatan 1 production run, for example. This is illustrated by the example of around 5µm. Hence the possibility
suspension analysis it can be
Chapman & Hall. 4th Edition, 1992. the basic problem of particle size equivalent spheres of cylinders for disputing answers that different
valuable in assessing the amount 750 15 UPPSALA
ISBN 04123570 analysis – how do we describe a (Fig. 2). However, if our cylinder techniques give!
of agglomerated material in the Sweden 3-dimensional object with one
3. G.J.J.Beckers, H.J.Veringa; Powder changes shape or size then the Note also that all these cylinders will
dry state. Tel: +46 (0) 18 55 24 55 number only? volume/weight will change and
Technology 60 (1989) p245-248. appear the same size to a sieve, of say
● Liquid suspensions and emul- Fax: +46 (0) 18 55 11 14 Figure 1 shows some grains of sand. we will at least be able to say that
4. J. Ranucci Pharmaceutical Technology; 25µm where it will be stated that "all
sions can be measured in a October 1992 pp 108-114. What size are they? it has got larger/smaller etc. with material is smaller than 25µm". With
recirculating cell and this gives www.malvern.co.uk our equivalent sphere model. laser diffraction these ‘cylinders’ will
5. G. Hind; Manufacturing Chemist August
high reproducibility and also 1990 pp 28-30. The equivalent sphere be seen to be different because they
MRK034 possess different values.
allows dispersing agents (e.g. 0.1% 6. M.W.Wedd; ILASS-Europe 8th Annual There is only one shape that can be
Calgon, sodium hexametaphos- conference, Koninklijke/Shell described by one unique number and 100µm 39µm
phate solution for TiO2 ) and Laboratorium, Amsterdam 30 September that is the sphere. If we say that we
surfactants to be employed to –2nd October, 1992. have a 50µ sphere, this describes it
ascertain the primary particle size. 7. T. Hatch & S.P Choate. J. Franklin Inst. exactly. We cannot do the same even
207 pp 369-387 (1929). for a cube where 50µ may refer to an Figure 2
If possible the preferred method 20µm

8 1
 This also means that there cannot be to get back to a mean diameter: These means are analagous to which distributions we use.
Size of cylinder
Aspect
Equivalent
anything like particle size standard moments of inertia and introduce X nl = D [1,0] = 1 + 2 + 3 = 2.00
2 3 Again neither distribution is incor-
Ratio
Sperical
for particles like grains of sand. (1 2+ 2 2+ 3 2) = 2.16 = ∑ d another linear term in diameter rect. The data are just being examined
Height Diam. Diameter 3 n
Standards must be spherical for (i.e. surface area has a d3 dependence in different ways. If we were making
20 20 1:1 22.9 comparison between techniques. and volume or mass a d4 dependence Xns= D [2,0] = 1 + 4 + 9 = 2.16 a space suit, for example, we could
However we can have a particle size This is again a number mean as below): 3 say that it is easy to avoid the 7000
40 20 2:1 28.8 standard for a particular technique (number-surface mean) because the 4 large objects and this takes care of
D [4,3] = 13+ 22+ 33 = 2.72 = ∑ d 3
4 4 4
100 20 5:1 39.1 and this should allow comparison number of particles appear on the 3 99.96% of all cases. However, what is
between instruments which use bottom of the equation. We have 1+2+3 ∑d Xnv= D [3,0] = 1 + 8 + 27 = 2.29
3 more important with a space suit is
200 20 10:1 49.3 that technique. summed the squares of the diameter the protection against small particles
400 20 20:1 62.1 so in mathematical terms this is 3 which are 99.3% by number!
D [3,2] = 12+ 22+ 32 = 2.57 = ∑ d 2
3 3 3
called the D[2,0] – diameter terms
10 20 0.5:1 18.2 D[4,3]etc
squared on the top, no diameter 1+2+3 ∑d X ls = D [2,1] = 1 + 4 + 9 = 2.33 If we take a calculator and calculate
1+2+3 the means of the above distributions
Imagine three spheres of diameters terms on the bottom.
4 20 0.2:1 13.4 we find that the number mean is
1,2,3 units. What is the average size These formulae indicate around
If I am a chemical engineer I will about 1.6cm and the mass mean
2 20 0.1:1 10.6 of these three spheres? On first which central point of the frequency 1 + 8 + 27 = 2.45
want to compare the spheres on the Xlv = D [3,1] = about 500cm – again very different.
reflection we may say 2.00. How the (surface area or volume/mass) 1+2+3
basis of weight. Remembering that
have we got this answer? We have distribution would rotate. They are,
Different techniques the weight of a sphere is:
summed all the diameters in effect, centres of gravity of the Interconversion between number,
Clearly if we look at our particle 4 π r 3 .p respective distributions.The advantage Xsv = D [3,2] = 1 + 8 + 27 = 2.57 length and volume/mass means.
under the microscope we are looking ( ∑ d = 1 + 2 + 3) 3 of this method of calculation is 1+4+9 If we are measuring our particles on
at some 2-D projection of it and obvious – the formulae do not an electron microscope we know,
there are a number of diameters that contain the numbers of particles and from an earlier section (Different
then we must cube the diameters, Xvm= D [4,3] = 1 + 16 + 81 = 2.72
we can measure to characterise our and divided by the number of parti- therefore calculations of the means techniques give different means.) that
divide by the number of particles and 1 + 8 + 27
particle. If we take the maximum cles (n=3). This is a number mean, and distributions do not require we are calculating the D[1,0] or the
take a cube root to get back to a
length of the particle and use this as (more accurately a number length knowledge of the number of number-length mean size. If what
mean diameter:
our size, then we are really saying that mean), because the number of the particles involved. Laser diffraction = X wm we really require is the mass or
our particle is a sphere of this maxi- particles appears in the equation: 3
(1 3+ 2 3+ 3 3) = 2.29 = ∑ d
3 initially calculates a distribution volume mean size we have to convert
mum dimension. Likewise, if we use Mean diameter = 3 n based around volume terms and this our number mean to a mass mean.
the minimum diameter or some other is why the D[4,3] is reported in a Number and volume distributions Mathematically, this is easily feasible,
quantity like Feret’s diameter, this will 1 + 2 + 3 = 2.00 = ∑ d Again this is a number mean
prominent manner. but let us examine the consequences
give us another answer as to the size 3 n of such a conversion.
of our particle. Hence we must be (number-volume or number-weight Size Number of % by % by
mean) because the number of Different techniques give Imagine that our electron
aware that each characterisation tech- In mathematical terms this is called different means. (cm) Objects Number Mass
particles appears in the equation. measurement technique is subject to
nique will measure a different proper- the D[1,0] because the diameter In mathematical terms this can be If we use an electron microscope to an error of ±3% on the mean size.
ty of a particle (max. length, min. terms on the top of the equation are 10-1000 7000 0.2 99.96
seen to be D[3,0]. measure our particles it is likely that When we convert the number mean
length, volume, surface area etc.) and to the power of (d1 ) and there are
The main problem with the simple we will measure the diameters with a 1-10 17500 0.5 0.03 size to a mass mean size then as the
therefore will give a different answer no diameter terms (d0 ) on the
means, D[1,0], D[2,0], D[3,0], is that graticule, add them up and divide by mass mean is a cubic function of the
from another technique which meas- bottom equation. 0.1-1.0 3500000 99.3 0.01
the number of particles is inherent in the number of particles to get a mean diameter then our errors will be
ures an alternative dimension. Figure
However imagine that I am a catalyst the formulae. This gives rise to the result. We can see that we are Total 3524500 100.00 100.00 cubed or ±27% variation on the
3 shows some of the different answers
engineer. I will want to compare need to count large numbers of generating the D[1,0] number-length final result.
possible for a single grain of sand.
these spheres on the basis of surface particles. Particle counting is mean by this technique. If we have However, if we are calculating the
Each technique is not wrong – they The above example is adapted from
area because the higher the surface normally only carried out when the access to some form of image analysis mass or volume distribution as we do
are all right – it is just that a different an article in New Scientist (13
area, the higher the activity of the numbers are very low (in the ppm or then the area of each particle is with laser diffraction then the
property of the particle is being meas- October 1991). There are a large
catalyst. The surface area of a sphere ppb regions) in applications such as measured and divided by the number situation is different. For a stable
ured. It is like you measure your number of man-made objects orbit-
is 4πr2. Therefore to compare on contamination, control and cleanli- of particles – the D[2,0] is generated. sample measured under recirculating
matchbox with a cm ruler and I ing the earth in space and scientists
basis of surface area we must square ness. A simple calculation shows that If we have a technique like conditions in liquid suspension, we
measure with an inch ruler (and you track them regularly. Scientists have
the diameters, divide by the number in 1g of silica (density 2.5) then there electrozone sensing, we will measure should be able to generate a volume
measure the length and I measure the also classified them in groups on the
of particles, and take the square root would be around 760 x 109 particles the volume of each particle and mean reproducibility of ±0.5%.
width!). Thus we can only seriously basis of their size.
if they were all 1µm in size. divide by the number of particles – If now we convert this volume mean
compare measurements on a powder If we examine the third column
a D[3,0] is generated. to a number mean the error or the
by using the same technique. Hence the concept of Moment above we would conclude (correctly)
Sphere of same
Means needs to be introduced and Laser diffraction can generate the number mean is the cube root of
minimum length that 99.3% of all particles are incredi-
Sphere of
this is usually where some confusion D[4,3] or equivalent volume mean. 0.5% or less than 1.0%!
dw
same weight bly small. This is evaluating the data
Sphere of same dmin can arise.The two most important This is identical to the weight equiv- In practice this means that if we are
maximum length dmax on a NUMBER basis. However, if
moment means are the following: alent mean if the density is constant. using an electron microscope and
we examine the fourth column we
dv Sphere of
same volume ● D[3,2] – Surface Area Moment So each technique is liable to gener- would conclude (correctly) that vir- what we really want is a volume or
Figure 3 ate a different mean diameter as well mass distribution, the effect of
Mean – Sauter Mean Diameter tually all the objects are between 10
as measuring different properties of – 1000cms. This is where all the ignoring or missing one 10µ particle
● D[4,3] – Volume or Mass
Sphere having same
dsed our particle. No wonder people get MASS of the object is. Note that the is the same as ignoring or missing
sedimentation rate Moment Mean – De Brouckere confused! There are also an infinite one thousand 1µ particles.Thus we
ds Mean Diameter NUMBER and MASS distribution
Sphere of same
surface area
number of "right" answers. Imagine are very different and we would draw must be aware of the great dangers
Sphere passing same
sieve aperture
dsieve 3 spheres with diameters 1,2,3 units: different conclusions depending on of interconversion.

2 3
On the Malvern Sizers both the diameters. For example, a micro- of the system lies. This is where the MODE ● D[3,2] is the surface area moment Sedimentation
DOS™ and Windows™ software will scope will measure the D[1,0] and D[4,3] is much more useful. mean or the Sauter Mean
This is the most common value of This has been the traditional method
calculate other derived diameters but will/may derive other diameters from In our two sphere example the mass Diameter (SMD).
the frequency distribution i.e. the of measurement in the paint and
we must be very careful of how we this. or volume moment mean would be highest point of the frequency curve. ceramics industry and gives
interpret these derived diameters. We can place more faith in the calculated as follows: Methods of measurement seductively low answers! The
Different means can be converted to measured diameter than we can on Normal or Gaussian Distribution
From our earlier sections, we have applicable range is 2-50 microns
each other by means of the following the derived diameters. In fact, in 4 4
D [4,3] = 13+ 10 3 = 9.991 Mean
seen that each measurement (Ref. 1 & 2) despite what the
equations (Hatch-Choate some instances it can be very 1 + 10 Median Mode technique produces a different answer manufacturers may claim.
transformation) (Ref.7): dangerous to rely on the derived
because it is measuring a different The principle of measurement is
property. For example, the Malvern This value shows us more where the based on the Stokes’ Law equation:
2 dimension of our particle. We will
InD4.3= 1lnX v + 0.5ln o analysis table gives us a specific mass of the system lies and is of more %
now discuss some of the relative
surface area in m 2/cc or m 2/gm. value to chemical process engineers. 2
(ps- pf ) D g
advantages and disadvantages of the Terminal Velocity, Us=
We must not take this value too However, let us imagine that we are main different methods employed.
InD4.2= 2lnX v literally – in fact, if what we really in a clean room making wafers of sil-
want is the specific surface area of icon or gallium arsenide. Here, if Diameter Equipment can be as simple as
our material we really should use a one particle lands on our wafer it Sieves the Andreason pipette or more
2
surface area specific technique e.g. Figure 4
InD4.1= 3lnX v - 1.5ln o will tend to produce a defect. In this This is an extremely old technique complicated involving the use of
B.E.T. or mercury porosimetry. instance the number or concentration Imagine that our distribution is a but has the advantage that it is cheap centrifuges or X-rays.
2 of the particles is very important Normal or Gaussian distribution. and is readily usable for large particles Examination of this equation will
InD4.0= 4lnX v - 4ln o because 1 particle = 1 defect. We The mean, median and mode will such as are found in mining.
Which number do we use? indicate one or two potential pitfalls.
would want to use a technique that lie in exactly the same positions. Terence Allen (Ref. 2) discussed the The density of the material is
Remembering that each different
2 directly measures the number of par- See Figure 4. difficulties of reproducible sieving needed, hence the method is no
InD3.2= 1lnX v - 0.5ln o technique measures a different
ticles or gives us the concentration of However, imagine that our but the main disadvantages to many good for emulsions where the
property (or size) of our particle and
particles. In essence this is the differ- distribution is bimodal as shown users are the following: material does not settle, or very dense
that we may use the data in a num-
2 ence between particle counting and in Figure 5. ● Not possible to measure sprays materials which settle quickly. The
InD3.1= 2lnX v - 2ln o ber of different ways to get a differ-
particle sizing. With counting we or emulsions end result is a Stokes diameter (DST)
ent mean result (D[4,3], D[3,2] etc.), Bimodal Distribution
will record each particle and count it which is not the same as a weight
then what number should we use? ● Measurement for dry powders
2 – the size is less important and we Mean
diameter, D[4,3], and is simply a
InD3.0= 3lnX v - 4.5ln o Let’s take a simple example of two may only require a limited number of Median
under 400# (38µ) very difficult.
spheres of diameters 1 and 10 units. Wet sieving is said to solve this comparison of the particle’s settling
size classes (say 8). With sizing the Mode
rate to a sphere settling at the same
Imagine that we are making gold. problem but results from this
2 absolute number of particles is less rate. The viscosity term in the
InD2.0= 2lnX v - 4ln o If we calculate the simple number % technique give very poor
relevant than the sizes or the size dis- denominator indicates that we will
mean diameter this will give us: 49% 1% 51% reproducibility and are difficult
tribution of the particles and we may need to control temperatures very
to carry out.
2 require more size bands. accurately – a 1°C change in
InD2.1= 1lnX v - 1.5ln o Diameter ● Cohesive and agglomerated
D [1.0] = 1 + 10 = 5.50 For a metered dose inhaler for asth- Figure 5
materials e.g. clays are difficult temperature will produce a 2%
2 ma sufferers then both the concentra- change in viscosity.
2 The mean diameter will be almost to measure.
InD1.0= 3lnX v - 2.5ln o tion of drug and its particle size dis- With the equation it is relatively easy
So we would assume that the average exactly between the two distributions ● Materials such as 0.3µ TiO2 are
tribution is important. to calculate settling times. It can be
size of the particles in the system is as shown. Note there are no particles simply impossible to measure and
5.50 units. However, we must which are this mean size! The resolve on a sieve. The method is shown that a one micron particle of
Measured and derived diameters. remember that if we are making gold Mean, Median and Mode – median diameter will lie 1% into the not inherently high resolution. SiO2 (ρ = 2.5) will take 3.5 hours to
We have seen that the Malvern laser we are interested in the weight of basic statistics higher of the two distributions settle 1cm under gravity in water at
● The longer the measurement, the
diffraction technique generates a our material. It is important to define these three because this is the point which 20°C. Measurements are therefore
smaller the answer as particles extremely slow and repeat
volume distribution for the analysed For example, if we have a process terms as they are so often misused divides the distribution exactly into orientate themselves to fall
light energy data. (Note that with stream we are not interested that in both statistics and particle two. The mode will lie at the top of measurements tedious. Hence the
through the sieve. This means that move to increase g and attempt to
Fraunhofer analysis, the projected there are 3.5 million particles in it, size analysis: the higher curve because this is the measurement times and operating
area distribution is assumed). This we are more interested that there is most common value of the diameter remedy the situation.
methods (e.g. tapping) need to be
volume distribution can be converted 1kg or 2kg of gold. (only just!). rigidly standardised. The disadvantages of increasing ‘g’
MEAN are discussed in (Ref. 3).
to any number or length diameter Remembering that the mass mean This example illustrates that there is
This is some arithmetic average of ● A true weight distribution is not
as shown above. is a cubic function of diameter, we no reason which the mean, median More specific criticisms of the
the data. There are a number of produced. Rather the method sedimentation technique are to
However, in any analysis technique, would see that the sphere of diameter and mode should be identical or even relies on measuring the second
means that can be calculated for similar, it depends on the symmetry be found in (Ref. 2).
we must be aware of the consequences 1 unit has a mass of 1 unit and the smallest dimension of the
particles (see section D[4,3] etc.). of the distribution. Stokes’ Law is only valid for spheres
of such a conversion (see previous sphere of diameter 10 units has a particle. This can give some
section "Interconversion between mass of 10 3 = 1000 units. Note that in the Malvern strange results with rod-like which possess the unique feature of
number, length and volume/mass That is, the larger sphere makes up
MEDIAN analysis table: materials e.g. paracetamol in the being the most compact shape for
means.") and also which mean 1000/1001 parts of the total mass of This is the value of the particle size pharmaceutical industry. the volume or surface area they
● D[4,3] is the volume or mass
diameter is actually measured by the the system. If we are making gold which divides the population exactly possess. Hence more irregularly
moment mean or the ● Tolerance. It is instructive to
shaped ‘normal’ particles will possess
equipment and which diameters are then we can throw away the sphere into two equal halves i.e. there is De Broucker mean. examine a table of ASTM or BS more surface area than the sphere
really calculated or derived from that of 1 unit because we will be losing 50% of the distribution above this
● D[v,0.5] is the volume (v) median sieve sizes and see the permitted and will therefore fall more slowly
first measured diameter. less than 0.1% of the total mass of the value and 50% below.
diameter sometimes shown as D50 tolerances on average and maxi- because of the increased drag than
Other techniques will generate other system. So the number mean does or D0.5. mum variation. The reader is
diameters from some measured not accurately reflect where the mass their equivalent spherical diameters.
invited to do this.
4 5
 Comparison of Brownian movement displacement and gravitational settling displacement ● Dense materials or large materials a very valuable aid to the through the beam by means of
are difficult to force through the characterization of particles. pressure and sucked into a vacuum
Displacement in 1.0 second (µm) orifice as they sediment before Laser diffraction cleaner to prevent dust being
this stage. sprayed into the environment.
In air at 70°F (1atm) In water at 70°F In water at 70°F This is more correctly called Low
So, in summary this technique is Particles in suspension can be
Particle Due to Due to Due to Due to  Angle Laser Light Scattering
k =100(%) excellent for blood cells but of a measured by recirculating the
diameter (µm) Brownian movement* Gravitational settling+ Brownian movement* Gravitational settling+ (LALLS). This method has become
more dubious nature for many sample in front of the laser beam.
0.10 29.4 1.73 2.36 0.005 31.1 96.9 the preferred standard in many
industrial materials. industries for characterization and Older instruments and some existing
0.25 14.2 6.3 1.49 0.0346 3.15 75.9
quality control. The applicable range instruments rely only on the
0.50 8.92 19.9 1.052 0.1384 0.556 35.7
Microscopy according to ISO13320 is 0.1 – Fraunhofer approximation
1.0 5.91 69.6 0.745 0.554 0.0983 5.0 which assumes:
This is an excellent technique as it 3000µm. Instrumentation has been
2.5 3.58 400 0.334 13.84 0.00995 1.0
allows one to directly look at the developed in this field over the last ● Particle is much larger than the
10.0 1.75 1550 0.236 55.4 0.00031 0.03 wavelength of light employed
particles in question. So the shape of twenty years or so. The method
*Mean displacement given by equation (7.20) the particles can be seen and it can relies on the fact that diffraction (ISO13320 defines this as being
+Distance settled by a sphere of density 2000kgm-3, including Cunningham’s correction. also be used to judge whether good angle is inversely proportional to greater than 40λ i.e. 25µm when
 is defined in equation (7.23) (Taken from Reference 2 p 259) particle size. a He-Ne laser is used).
dispersion has been achieved or
whether agglomeration is present in Instruments consist of: ● All sizes of particle scatter with
For objects like kaolins which are and the system is inaccurate. can be corrected for by measuring
disc-shaped this effect is even more the system. The method is relatively equal efficiencies.
Above 50µm, settling is turbulent the area under the peak rather than ● A laser as a source of coherent
accentuated and large deviations and Stokes’ Law again is not the peak height. For blood cells the cheap and for some microscope intense light of fixed wavelength. ● Particle is opaque and transmits
from reality are to be expected. applicable. technique is unsurpassed and the systems it is possible to use image no light.
He-Ne gas lasers (λ=0.63µm) are
Furthermore, with small particles method is capable of giving both a analysis to obtain lists of numbers the most common as they offer These assumptions are never correct
there are two competing processes – Figure 6 shows the expected
number count and volume (usually to 6 or 8 places of decimals, the best stability (especially with for many materials and for small
gravitational settling and Brownian differences between a sedimentation
distribution. For real, industrial well beyond the resolution of respect to temperature) and better material they can give rise to errors
motion. Stokes’ Law only applies to and laser diffraction results.
materials such as pigments there are a the technique!). signal to noise than the higher approaching 30% especially when
gravitational settling. The table at the Kaolin
top of this page shows a comparison number of fundamental drawbacks: It is interesting to note that 1g of wavelength laser diodes. It is to the relative refractive index of the
between the two competing Sedimentation ● Difficult to measure emulsions. 10µm particles (density 2.5) contains be expected when smaller laser material and medium is close to
processes. It will be seen that very (Sprays not possible!) Dry 760 x 10 6 particles – all these can diodes can reach 600nm and unity. When the particle size
large errors (approx. 20%) will result % Laser
powders need to be suspended never be examined individually below and become more reliable approaches the wavelength of light
if sedimentation is used for particles by microscopy. that these will begin to replace the the scattering becomes a complex
under 2µm in size and the errors in a medium so cannot be
measured directly. However, it is not suitable as a quality bulkier gas lasers. function with maxima and minima
will be in excess of 100% for
0.5µm particles. or production control technique ● A suitable detector. Usually this is present. The latest instruments
● Must measure in an electrolyte.
1 2 5 10 100
beyond a simple judgement of the a slice of photosensitive silicon (e.g. Mastersizer 2000, Malvern
The sedimentation technique gives Size For organic based materials this is
an answer smaller than reality and Figure 6 type indicated above. Relatively few with a number of discrete Instruments) use the full Mie theory
difficult as it is not possible to
this is why some manufacturers particles are examined and there is detectors. It can be shown that which completely solves the equa-
measure in xylene, butanol and
deceive themselves. In summary the Electrozone sensing the real danger of unrepresentative there is an optimum number of tions for interaction of light with
other poorly conducting solutions.
main disadvantages of the technique (Coulter Counter) sampling. Furthermore, if a weight detectors (16-32) – increased matter. This allows completely accu-
for pigment users are the following: ● The method requires calibration rate results over a large size range
This technique was developed in distribution is measured the errors numbers do not mean increased
● Speed of measurement. Average standards which are expensive and (0.02 -2000µm typically). The Mie
the mid 1950’s for sizing blood cells are magnified – missing or ignoring resolution. For the photon
times are 25 minutes to 1 hour for change their size in distilled water theory assumes the volume of the
which are virtually monomodal one 10µm particle has the same correlation spectroscopy technique
measurement making repeat and electrolyte (Ref. 2). particle as opposed to Fraunhofer
suspension in a dilute electrolyte. effect as ignoring one thousand (PCS) used in the range 1nm –
analyses difficult and increasing ● For materials of relatively wide 1µm particles. 1µm approximately, the intensity which is a projected area prediction.
The principle of operation is very particle size distribution the
the chances for reagglomeration. of light scattered is so low that a The "penalty" for this complete
simple. A glass vessel has a hole or Electron microscopy has elaborate
● Accurate temperature control. method is slow as orifices have to photomultiplier tube, together accuracy is that the refractive indices
orifice in it. Dilute suspension is sample preparation and is slow.
Needed to prevent temperature be changed and there is a danger with a signal correlator is needed for the material and medium need to
made to flow through this orifice With manual microscopy few
gradients and viscosity changes. of blocking the smaller orifices. to make sense of the information. be known and the absorption part
and a voltage applied across it. As particles are examined (maybe 2000
Inability to handle mixtures of dif- ● The bottom limit of the method in a day with a good operator) and ● Some means of passing the sample of the refractive index known or
● particles flow through the orifice the
fering densities – many pigments is set by the smallest orifice there is rapid operator fatigue. through the laser beam. In guessed. However, for the majority
capacitance alters and this is indicated
are a mixture of colouring matter available and it is not easy to Again there is the problem of practice it is possible to measure of users this will present no
by a voltage pulse or spike. In older
and extender/filler. measure below 2µm or so. "which dimension do we measure?" aerosol sprays directly by spraying problems as these values are either
instruments the peak height was
Certainly it is not possible to Hence there can be large operator them through the beam. This generally known or can be measured.
● Use of X-rays. Some systems use measured and related to a peak
measure TiO2 at 0.2µm. to operator variability on the same makes a traditionally difficult
X-rays and, in theory, personnel height of a standard latex. Hence the
should be monitored. method is not an absolute one but is ● Porous particles give significant sample. In combination with measurement extremely simple.
of a comparative nature. Problems of errors as the "envelope" of the diffraction microscopy becomes A dry powder can be blown
● Limited range. Below 2µm,
particle orientation through the beam particle is measured.
Brownian motion predominates

6 7